Hematogenous Splenic Metastases as an Independent Negative Prognosis Factor at the Moment of Primary Cytoreduction in Advanced Stage Epithelial Ovarian Cancer – A Single Center Experience

Discussion

Ovarian cancer is still a significant global health problem, being responsible for a high number of deaths among women worldwide annually (1, 2). The main patterns of spread are via the peritoneal route, inducing diffuse surface involvement of the abdominal viscera, the hematogenous route giving rise to parenchymatous visceral metastases, and in a lower proportion of patients, the lymphatic route. Once tumor dissemination goes beyond the pelvic area, upper abdominal involvement is frequently seen. In these cases, at primary cytoreduction, associated visceral resection, such as peritoneal diaphragmatic stripping, full-thickness diaphragmatic resection, splenectomy, and minor or even major hepatectomy have been proposed in order to achieve an R0 resection (2-4).

Although the efficacy of complete cytoreduction has been widely accepted and demonstrated by large series of cases, the benefits of extended resection in the upper abdomen was from the outset a matter of debate (3-6). Some authors considered that the presence of an upper abdomen tumor burden reflects a tumor with a more aggressive biological behavior which is secondarily associated with poorer outcome. However, in time, due to the formation of multidisciplinary teams involving visceral surgeons, hepato-bilio-pancreatic surgeons and gynecological oncologists, upper abdominal resections have become more frequent in order to obtain complete resection and have been proven to improve overall survival. Since postoperative outcomes are strongly related to the operative technique and to the postoperative management, and only in a low proportion to aggressive biological tumor behavior, upper abdominal resections have been widely used as a basic component of the therapeutic armamentarium against advanced-stage EOC (5-7, 12-16). Therefore, the preferred therapeutic option for patients with advanced-stage EOC with upper abdominal tumor burden remains an aggressive surgical approach followed by taxane- and platinum-based chemotherapy.

In one of the largest studies which focused on the importance of associated visceral resections at primary cytoreductive surgery, Eisenkop et al. included 203 cases and demonstrated that other therapeutic procedures such as diaphragmatic or bowel resection, or extended lymph node removal can be safely performed with no significant modification of the morbidity rate in order to achieve an R0 resection (16).

Parenchymatous liver metastases are usually associated with a poorer outcome, being considered as an inclusion criteria for FIGO stage IV ovarian cancer (8-10). Although it can be considered that parenchymatous involvement is indicative of poor prognosis, both hepatic and splenic metastases, a difference between the outcomes associated with the two hematogenous visceral involvements has been suggested. While in cases diagnosed with splenic metastases an R0 resection is easily achieved by complete resection of the involved organ, in cases presenting liver metastases, microscopic residual tumors might by still present in the remnant liver after complete macroscopic resection (10).

Splenic parenchymatous involvement is not a common condition, being reported in up to 41% of cases submitted to primary cytoreduction for advanced-stage EOC (16-19). Due to the limited number of cases presenting splenic involvement and to the various models of involvement (peritoneal, parenchymatous or even lymphatic seeding), initial studies reported conflicting results, failing to demonstrate if the presence of splenic involvement is itself a poor prognosis factor (17, 20, 21).

When estimating the risks and benefits of splenectomy as part of primary cytoreductive surgery, there are also only few studies focusing on this issue. In the study conducted by Eisenkop et al., 404 patients with advanced-stage EOC were included; at primary cytoreduction, complete resection was achieved in 356 patients, 49 of them also requiring splenectomy in order to provide R0 resection. Postoperative morbidity was non-significantly higher in the subgroup of patients submitted to splenectomy; however, in this subgroup, a higher number of pancreatic leaks was seen (6.1%), all of them being managed conservatively. When estimating the long-term results, patients in whom splenectomy was required had a median overall survival of 56.4 months and an estimated 5-year overall survival of 48%, comparable with the cases in whom splenectomy had not been performed; in the latter subgroup, the median overall survival was 76.8 months, while the estimated 5-year overall survival was 58% (p=0.4) (15).

In cases presenting liver metastases, the benefit of surgery has been widely demonstrated; different outcomes in terms of survival have been recorded for cases presenting parenchymatous versus peritoneal seeding (12-14). These aspects led to the hypothesis that similar outcomes might also be associated with splenic metastases and encouraged surgeons to perform similar studies regarding the spleen and the various patterns of spread. In our present study, the proportion of patients with parenchymatous splenic metastases was 28.7% at the time of primary cytoreduction, similar to that reported by other authors. In the study conducted by Tanner et al., overall incidence of parenchymatous splenic metastases of 21% was reported, while in a data review conducted by the same authors, an aggregate rate of hematogenous splenic involvement of 18% was provided (10).

In our cohort, a significant difference in terms of overall survival was also obtained for cases presenting parenchymatous involvement when compared to other patterns of spread (24.5 months versus 40.6 months for cases with hilar involvement and 58.4 months for cases with peritoneal seeding). These findings are in concordance with other literature data; for example, in a similar study conducted by Tanner et al. patients with parenchymatous splenic metastases had an overall survival rate of 28.5 months versus 51.2 months for cases with peritoneal seeding (p=0.004) (10). In the same study, other prognostic factors which significantly impacted on disease-free survival and overall survival were age, preoperative albumin level, FIGO stage (III versus IV), serous histology, the presence of bulky upper abdominal disease, and administration of intraperitoneal chemotherapy. In multivariate analysis, only age, preoperative albumin, residual disease, FIGO stage IV, and the presence of parenchymatous splenic involvement were independently associated with overall survival (10).

In another similar study conducted by Ayhan et al., a decrease of overall survival of 12 months was observed between patients with parenchymatous compared with those with peritoneal seeding, however, this was not statistically significant (20).

Contrarily to these results, in Magtibay et al.'s study involving patients submitted to splenectomy as part of primary and secondary cytoreductive surgery, no difference in terms of survival was reported between cases with parenchymatous versus those with peritoneal seeding. The study included 112 patients who underwent splenectomy for advanced-stage EOC; in 66 cases, splenic involvement was found at primary cytoreduction, while in the other 46 cases, splenectomy was performed for recurrent EOC. The overall median survival for the primary cytoreduction group was 21.6 months, with an estimated 2-year survival of 46.2%, while the overall median survival in the secondary cytoreduction group was 20.3 months, with an estimated 2-year survival of 42.3%. Neither at the time of primary nor secondary cytoreduction did the presence of parenchymatous splenic involvement have any significant impact on overall survival (p=0.96 and p=0.25, respectively). However, this might be explained by the small number of patients diagnosed with parenchymatous involvement (seven cases in the primary cytoreduction group and 11 cases in the secondary cytoreduction group, respectively) (22).

The current series demonstrate that splenectomy can be safely performed as part of primary cytoreduction. Although a morbidity rate of 30% and a mortality rate of 7% were encountered, these were not solely the result of complications related to splenectomy. Almost half of them seemed to be related to other performed surgeries, such as enteral resection, full-thickness diaphragmatic resection or to the preoperative associated co-morbidities, such as cardiac or respiratory disorders. Although a more significant benefit in terms of survival was achieved in patients with peritoneal or hilar involvement, cases retrospectively diagnosed with parenchymatous involvement also experienced a benefit in terms of survival.